Are Private and Government Jobs Data on the Same Page?

The discussion of non-traditional or “alternative” economic or financial data has increased in recent years, especially after the longest government shutdown last year that disrupted its data collection. Many of the alternative datasets are from the private sector as advances in technology have made it easier for organizations to access the data, which may offer more depth and a higher frequency of reporting relative to some government data.

Whether it is a pandemic, government shutdown, or a normal economic environment, the Federal Reserve, market participants, and various private and public organizations having access to multiple data sets may offer more insight to economic analysis for decision-making and help to complement government data.

Jobs data can also give some insight towards the strength of consumer spending, which accounts for about two-thirds of U.S. economic activity and offer a glimpse into future corporate earnings and the overall economic landscape.1 This in turn may have an impact on the capital markets. For example, an expanding labor market is a sign of a growing economy, which could impact equities, while any overheating of the economy, however, could cause the Federal Reserve to sustain or increase interest rates, which could affect the U.S. dollar’s value. Since many commodities are quoted in dollars (USD), an appreciating USD could make commodities cost more around the world, likely dampening demand. Higher interest rates could also cause higher borrowing costs for organizations and individuals.

Often the firms producing alternative data are not in business to produce data but are in a business related to the data. For example, such industries may include transportation, payroll processing, real estate, labor, and retail. It is their respective industry that allows them to accumulate data they specialize in. Alternative data may be available at a higher frequency or closer to “real-time” than government data.

In the U.S., the frequency of the search terms “alternative data” and “private data” grew in 2025. However, these terms experienced a reduction in searches in the U.S. (Fig. 1) after the 2025 federal government shutdown ended.

Source: https://trends.google.com/trends/

When thinking about alternative data, the ADP’s National Employment Report (NER) data release date is often listed on economic calendars and is a focus of the financial press. Therefore, it begs the question, what is ADP’s NER data and how does it compare with the Bureau of Labor Statistics (BLS) Nonfarm Payroll (NFP) monthly updates of employment data?

ADP is a global human capital management company that includes payroll processing services gathering data for more than 500,000 firms and over 26 million employees. This equates to about 20% of U.S. private employment.3 The monthly NER is produced by the ADP Research Institute division of ADP in collaboration with the Stanford Digital Economy Lab. The aggregated data covers non-farm private employment.4

The BLS’ monthly NFP data is usually released the first Friday of each month. The ADP data is released two days before the BLS data and may offer some insight to the condition of the private sector jobs market. Ultimately, both datasets are asking the same question, is the U.S. economy experiencing job expansion or contraction?

One can argue BLS’ NFP data, and ADP’s NER data are similar, with some differences. ADP’s data is derived from private firms that are within their customer universe and includes small to large firms but skewed towards mid-size to large firms. The NFP data includes about 121,000 private and government agencies ranging from small to large. Firms with less than 20 employees equate to about 45% of the BLS establishment survey.5

Both datasets offer monthly data, however, the timing and methodologies may vary as ADP is based on aggregated real-time payroll processing data of their customers. ADP will revise their data once to capture active employees who are paid monthly and are in their data after the reference week for a respective month.6 Both datasets use the 12th of the month as part of their reference week.7

The NFP data is derived from surveys of organizations and statistical modeling. Each report has the potential for two revisions within the two following months as some respondents submit their survey after the release date. The BLS also calculates an annual benchmark revision. The first benchmark revision is an estimate in September, and a final revision is released in February for the previous year to reduce statistical errors.8

Once BLS chooses a firm to survey, they are surveyed for at least two years. The monthly survey includes questions about “how many civilians were on their payroll during the pay period that included the 12th day of the month.” The employers must respond for two months in a row to be included in the month-to-month employment change estimates.9

Over the longer term, the BLS and ADP data tend to trend in the same direction (aka co-movement) even though the monthly data may be derived from different methods and different parts of the labor market.

Figures 2 and 2A show the overall levels of BLS and ADP since 2010 (ADP data inception). BLS has two datasets, one with private and government jobs- this is the metric usually reported in the press; and a second for private jobs only. The monthly changes between NFP and ADP are more likely to vary because government jobs may add and reduce positions on different cycles from the private market. For example, during recessions or economic slowdowns, there could be more hiring in government agencies. This divergence of data could offer investors to seek a “flight to quality” such as U.S. Treasuries. Or government policy changes may occur at various points of an economic cycle reducing budgets at the local, state or federal government levels.

The NFP Total captures, on average, around 22 million additional workers each month than ADP when one includes public employees (Fig. 2).

The gap between the measured levels of employment is “only” 1.257 million on average when comparing the BLS private sector payrolls to ADP which offers a more logical “apples to apples” comparison between the two series (Fig. 2A).

Except for COVID-19, the index data suggests (Fig. 2 & 2A) job creation remained relatively constant from 2010 until the spring of 2025. However, since the spring of 2025, the job data is suggesting a flattening or slowing labor market

Source: Bloomberg Professional (NFP T Index, ADP LEVL Index)

Source: Bloomberg Professional (NFP P Index, ADP LEVL Index)

Figure 2B shows monthly changes of the NFP private sector data and ADP data. However, due to the large job losses during COVID-19 in 2020, the data compresses the chart, therefore, it’s better to view the monthly changes without the pandemic period (Fig. 3 & 3B).

Source: Bloomberg Professional (NFP TCH Index, ADP CHNG Index)

Source: Bloomberg Professional (ADP CHNG Index, NFP PCH Index)

Source: Bloomberg Professional (ADP CHNG Index, NFP PCH Index), CME Group Economic Calculations

The difference between the monthly changes of NFP Total vs ADP tends to be larger than NFP Private to ADP with the average difference of 31,000, a median of 19,000 and standard deviation of 258,000, versus the NFP Private data with an average difference of 19,000, a median of 7,000, and a standard deviation of 228,000. (Fig. 2C).10

Source: Bloomberg Professional (NFP TCH Index, ADP CHNG Index, NFP PCH Index), CME Group Economics Calculations

As the U.S. economy came out of the financial crisis, the one-year moving averages for the NFP Private and ADP data were similar. The one-year average number of job changes in the private sector peaked in the 2014/ 2015 period around 250,000 and remained around 200,000 with a gradual decline over the next several years. In the second half of 2019 the ADP one-year moving average diverged from the NFP Private one-year moving average (Fig. 3).

As the U.S. economy recovered from COVID, the data notes large monthly job increases. However, monthly job increases peaked in the summer of 2021 with a one-year average peaking in the summer of 2022. By the end of 2024, job creation in the private sector was relatively consistent with a one-year moving average of 130,000 to 145,000 jobs per month. The data in both datasets became more mixed in the summer and fall of 2025 with December showing a one-year average around 52,000 to 62,000 net new jobs per month.

Viewing the rolling correlations of NFP Private to ADP can vary depending on the rolling period (Fig. 4). For example, a six-month rolling correlation shows a lot of correlation variance as it cycles between positive and negative correlations. This points out the monthly data can vary between the two datasets. The five-year rolling correlation removes the noise and tends to sustain a relatively consistent positive correlation over the longer term.

Source: Bloomberg Professional (ADP CHNG Index, NFP PCH Index), CME Group Economic Calculations

Viewing the rolling correlations in box and whisker plots (Fig. 5) shows rolling correlation ranges and respective averages (X) and medians (line) for each timeseries. The longer the rolling correlation, the higher the average and median and the smaller the variance. The rolling correlation periods at 0.99 occurred in April 2020 due to the historically large number of firms reducing their labor force during COVID-19.

The box and whisker plots suggest the correlations tend to be higher with the NFP Private correlation to ADP versus NFP Total correlation to ADP (Fig. 5 & 5A).

Source: Bloomberg Professional (NFP TCH Index, ADP CHNG Index), CME Group Economics Calculations

Source: Bloomberg Professional (NFP TCH Index, ADP CHNG Index), CME Group Economics Calculations

The BLS’ NFP monthly data can vary from the ADP’s NER monthly data on a month-to-month basis, but when you compare the datasets over the longer term, they tend to generally move in the same direction and complement each other by giving a broader view of the labor market. Comparing the rolling correlations suggest the correlations increase when comparing the NFP’s private sector data to the ADP versus the rolling correlation of the NFP’s Total dataset.

Ultimately, both datasets are asking the same question – is the U.S. economy experiencing job expansion or contraction? This data can be paired with other employment metrics such as the U3 or U6 unemployment rates and offer a complementary economic insight about interest rates, the Fed, equities, commodities and the forex markets.

https://fred.stlouisfed.org/series/DPCERE1Q156NBEA

Source: https://trends.google.com/trends/ “Numbers represent search interest relative to the highest point on the chart for the given region and time. A value of 100 is the peak popularity for the term. A value of 50 means that the term is half as popular. A score of 0 means there was not enough data for this term.”

https://adpemploymentreport.com/

https://www.bls.gov/news.release/pdf/empsit.pdf

https://www.adpresearch.com/how-representative-is-adp-employment-data/#:~:text=ADP%20payroll%20data%20also%20has,policy%20and%20business%20decision%2Dmaking

https://www.adpresearch.com/when-the-economic-data-is-confusing-take-a-closer-look/

https://www.bls.gov/news.release/prebmk.nr0.htm

https://www.everycrsreport.com/reports/IF13084.html

Due to the large job losses during COVID, Figure 2C excludes the changes from April, May, and June of 2020.

Navigate dynamic global markets with unparalleled price discovery and around-the-clock liquidity.

Find out more Mark Shore Economic Research Education Futures Article Equity Index E-mini Dow Micro E-mini Dow E-mini Russell 2000 E-mini S&P MidCap 400

All examples in this report are hypothetical interpretations of situations and are used for explanation purposes only. The views in this report reflect solely those of the author and not necessarily those of CME Group or its affiliated institutions. This report and the information herein should not be considered investment advice or the results of actual market experience.